Multiple invalidities?

John White

The theory of multiple intelligences has been influential in school reform across the
world. In England, for instance, it is widely used to back the idea that pupils have
preferred ‘learning styles’: some make better progress if they can involve their musical or
interpersonal or other strengths in their learning than if they have to be dependent on
language ability alone.

But does MI theory hold water?

Everything turns on the claim that there are a few relatively discrete intelligences:
linguistic, musical, logico-mathematical, spatial, bodily-kinaesthetic, intrapersonal and
interpersonal, to which have now been added naturalist and possibly existential
intelligences. One reason for the popularity of MI theory is its rejection of the unitary
general intelligence associated with IQ testing. Children who have been seen, or have
seen themselves, as dim are recognized to have other strengths. This is an important
thought. But it could be true and MI theory false. Long ago the philosopher Gilbert Ryle
(1949:48) reminded us that ‘the boxer, the surgeon, the poet and the salesman’ engage in
their own kinds of intelligent operation, applying ‘their special criteria to the performance
of their special tasks’. On his view, intelligent action has to do with flexible adaptation of
means in pursuit of one’s goals. This means that there are as many types of human
intelligence as there are types of human goal. Gardner has corralled this variousness into
a small number of categories. Is this justified?

Everything turns on how the intelligences are identified. The basic text here is Gardner
1983, ch 4. He writes

First of all, what are the prerequisites for an intelligence: that is, what are the
general desiderata to which a set of intellectual skills ought to conform before that
set is worth consideration in the master list of intellectual competences? Second,
what are the actual criteria by which we can judge whether a candidate
competence, which has passed the “first cut”, ought to be invited to join our
charmed circle of intelligences? (p60)

1
Identifying an intelligence is thus a two-stage process.

Prerequisites

The first stage is, in a way, the more important. If a candidate fails here, it stands no
chance. So what Gardner says about prerequisites is crucial. He tells us (1983:60-1) that

A human intellectual competence must entail a set of skills of problem-solving…

and must also entail the potential for finding or creating problems – thereby laying
the groundwork for the acquisition of new knowledge. These prerequisites
represent my effort to focus on those intellectual strengths that prove of some
importance within a cultural context.

He goes on to say (p62) that

a prerequisite for a theory of multiple intelligences, as a whole, is that it captures a

reasonably complete gamut of the kinds of abilities valued by human cultures.

Failing candidates

Which candidates fail and which pass the test? Among failures, Gardner includes the
‘ability to recognize faces’ and the ‘abilities used by a scientist, a religious leader, or a
politician’ (p61). The former ‘does not seem highly valued by cultures’. The latter
abilities are of great importance,

yet, because these cultural roles can (by hypothesis) be broken down into
collections of particular intellectual competences, they do not themselves qualify
as intelligences. (ibid)

Is it true that the ability to recognize faces is not valued by cultures? This seems
counterintuitive. For if most of us could not recognize the faces of our relatives, friends,
colleagues, or political leaders, it is hard to see how social life would be possible

How can one tell whether an ability is culturally important? Gardner writes as if there are
clear tests at this first of the two filters. Yet his very first example of a failure is
disputable.

2
His treatment of the ‘abilities used by a scientist, a religious leader, or a politician’ shows
that not only does an ability have to be valued within cultures: it has also not to be
breakable-down into more basic intellectual abilities. Presumably Gardner has in mind
some or all of the seven intelligences which he goes on to describe.

But this is also problematic. In what sense is being a political leader breakable down into
some or all of these seven? ‘Breaking down’ could hardly refer to sufficient, as distinct
from necessary, conditions. No combination of the intelligences is sufficient to produce
competence in political leadership, for otherwise those many of us who possess all the
seven (or so) intelligences to some degree would also possess this competence – and not
all of us do.

But if it is necessary conditions that are intended, while it is true that one cannot be a
political leader without linguistic competence or interpersonal understanding, it is also
true that one could not have logico-mathematical abilities without having some linguistic
abilities. This would cause a problem for Gardner, because if an intelligence cannot be
broken down - in the present sense - into other more basic abilities, then logico-
mathematical intelligence fails at the first hurdle, just like political competence.

Passing candidates

I turn now to those candidates which pass the first test. In Gardner 1983 these must
include the seven intelligences. They must have all been picked out for their problem-
solving and problem-creating skills important in human cultures.

Why these areas? It seems an odd list. For one thing, many of the items seem to be
logically interrelated – and this casts in doubt that these are ‘relatively autonomous’
competences (1983: 8). Mathematical abilities are, to a large extent, a specialized kind of
linguistic ability. One could not understand oneself and identify others, as the personal
competences require, unless one had some linguistic abilities. Logical abilities are not
necessarily tied to mathematical abilities, but are required by linguistic ability. There
could be no linguistic abilities unless people by and large possessed the spatial ability of
‘perceiving the visual world accurately’ (1983: 173). And so on.

Secondly, if one asks ‘what kinds of problem-solving and problem-creating competences

are valued in human societies?’, what sort of answer might be expected? There would
need to be more precision about the range of societies in question. Are we talking about
all of them, most, or only some of them? Gardner is not clear on this. On the one hand he
says

3
 The prerequisites are a way of ensuring that a human intelligence must be
genuinely useful and important, at least in certain cultural settings. (1983: 61)

On the other, Frames of Mind, like the wider Harvard Project of which it was the first
fruit, is a study of ‘human potential’ tout court, not human potential within certain
cultures. In a later work, Gardner has written

The theory is an account of human cognition in its fullness – I put forth the
intelligences as a new definition of human nature, cognitively speaking.. (Gardner
1999a: 44).

This would seem to suggest that the first filtering must be based on what all – or at least
virtually all - human societies value. How would we find out what this might be? We
have historical evidence stretching back a few millennia; and very patchy archeological
evidence taking us back another few. From these sources, and from what we know of the
kinds of creatures we are, it is reasonable to suppose that all, or nearly all human societies
have attached importance to linguistic ability, to some degree of interpersonal
understanding, to using parts of the body, to accurate visual perception (I am taking for
granted the logical interconnexions here). Have counting skills also been universally
important? Probably. About musical skills, we have less reason to be confident: even if all
known societies have prized these, what can we say about unknown ones? About
intrapersonal skills, we have every reason to be sceptical. These have to do with ‘access
to one’s own feeling life’ (1983: 239). We know this has been valued in certain
civilizations, eg in Athens, in Taoism, in the individualistic world of post-Renascence
times. But we have no grounds to expect to find it prized in the hypertraditional societies
that have been so prominent among known human cultures.

So the ‘first cut’ selection of the seven original intelligences looks as though it may have
been based on something other than a study of skills which all or nearly all human
societies have valued. I will come back to this later. In addition, there are other skills, not
included among the intelligences, which have as much prima facie plausibility for this
title as many so included: food-producing skills, for instance, shelter-building skills,
medical skills, child-rearing skills. Has Gardner considered these, but rejected them at the
‘second cut’, ie at the criteria stage?

It could be that he did not consider them at all, since he tells us that he understands
human intelligences as ‘human intellectual competences’ (my italics) and the skills just
mentioned are practical rather than intellectual. On the other hand, it is uncertain what his
term ‘intellectual’ covers. Gardner aside, in one common usage it has to do with activities
concerned with the pursuit of truth, like history, science, philosophy. More broadly, it can

4
also include artistic pursuits and perhaps the application of academic thinking to such
fields as politics and medicine. Gardner’s usage seems to go even wider. His bodily-
kinaesthetic intelligence covers the abilities of artisans, ball-players and instrumentalists
‘who are able to manipulate objects with finesse’. On this definition, good plumbers,
electricians, tilers, bricklayers come out as intellectuals.

On such a wide definition, perhaps human-wide skills to do with food-producing and

with shelter-building would be covered by Gardner’s scheme. It is not clear. All we can
say is that Gardner does not approach his ‘first cut’ via a comprehensive consideration of
what the valued problem-solving skills in any human society might be, drawing on
whatever empirical data is available. I will say more later about the approach he does
adopt. Meanwhile, I will leave problems thrown up by the ‘prerequisites’ condition, and
turn to the ‘criteria’.

Criteria

Once a candidate intelligence has satisfied the prerequisites, it has to meet various
criteria. These comprise (1983:62-9):

• potential isolation of the area by brain damage

• the existence in it of idiots savants, prodigies and other exceptional individuals

• an identifiable core operation/set of operations

• a distinctive developmental history, along with a definable set of expert 'end-state'

performances

• an evolutionary history and evolutionary plausibility

• support from experimental psychological tasks

• support from psychometric findings

• susceptibility to encoding in a symbol system.

There are problems about several of these items taken individually; as well as more

5
general problems about the identification and application of the criteria. I begin with
specific items. For convenience, I begin with two of them taken together.

‘an identifiable core operation/set of operations’

‘A distinctive developmental history, along with a definable set of expert 'end-state'

performances’

The interconnectedness of these two can be illustrated by reference to linguistic

intelligence. This has as its 'core operations' a sensitivity to the meaning of words, to
order among words, to the sounds and rhythms of words, and to the different functions of
language (1983:77). These core operations are seen at work 'with special clarity' in the
work of the poet.

Linguistic intelligence also possesses a distinctive developmental history, culminating in

expert ‘end-state’ performances like those of the poet. Syntactical and phonological
processes lie close to the core, since they unfold ‘with relatively scant need for support
from environmental factors’ (1983: 80-1). Other intelligences illustrate the same point.
Musical intelligence involves, as core operations, pitch, rhythm, and timbre (1983: 104-
5). It begins in infancy with rudimentary singing (108) and develops towards end-states
exemplified this time by composers. Spatial intelligence develops from such core abilities
as perceiving the visual world accurately, performing transformations on one's visual
experience, and recreating aspects of the latter (173). The expert end-state performances
are painting, sculpture and the sciences. Similar claims are made about the remaining
intelligences.

Gardner's theory of intelligence is developmentalist. Developmentalism is the theory that

the biological unfolding between two poles from seed through to mature specimen that
we find in the physical world - e.g. of plants, or human bodies - is also found in the
mental world. In his criteria, Gardner acknowledges the two poles in the mental case. At
one end, there are allegedly genetically given capacities common to human beings like
visual perception, innate knowledge of the rules of language (following Chomsky p.80),
the ability to move our bodies in different ways etc. At the other end is the mature state,
the 'definable set of expert "end-state" performances' mentioned among the criteria. We
have already seen examples in the highest flights of poetry, music, painting, sculpture and
science. Intrapersonal intelligence, whose core capacity or mental seed is 'access to one's
own feeling life', finds its full development in the work of a novelist like Proust or the
patient or therapist 'who comes to attain a deep knowledge of his feeling life' (1983:239).
Interpersonal intelligence, arising out of the primitive 'ability to notice and make
distinctions among other individuals' generates its 'highly developed forms...in political

6
and religious leaders (a Mahatma Gandhi or a Lyndon Johnson), in skilled parents and
teachers' etc. (ibid).

Problems in developmentalism

Gardner's theory faces an objection besetting all forms of developmentalism. This theory
is based on the assumption that the unfolding familiar in the biological realm is also
found in the mental.There are two problems about this, one for each of the two poles,
problems we see illustrated in Gardner's own writings.

i) First the seed, or initial state. What is characteristic of biological seeds, including the
union of sperm and egg at the beginning of human development, is that they have within
them the power to unfold into more complex stages, given appropriate environmental
conditions like air, light and water in the growth of a plant. To locate a parallel initial
state in the mental case it is not enough to pick out innately given capacities. There is no
doubt that such capacities exist. We are all born with the power to see and hear things, to
move our bodies, to desire food and drink, to feel certain basic emotions like fear, to feel
pain and pleasurable sensations. But we should not assume that any of the abilities just
mentioned have within them the power to unfold into more complex forms of the same
thing. I have italicized the word 'unfold' advisedly. For there is equally no doubt that
many of the primitive capacities just mentioned do change into more sophisticated
versions: the desire for food, for instance, becomes differentiated into desires for
hamburgers and ice-cream; the brute ability to move one's limbs becomes specified into,
for instance, the ability to run half-marathons or to tango. The changes wrought in these
capacities are cultural products: people are socialized into them. This social shaping
cannot be characterized, as a defence of developmentalism might urge, as environmental
conditions which have to be satisfied for natural processes of growth to occur - the
mental equivalents of air, light and water in the growth of plants. For in the latter case air,
light and water are causally necessary for the innate propensities to unfold from within,
while the cultural induction required in the case of learning language and other skills
dependent on this is logically necessary. Linguistic concepts cannot conceivably be
acquired on one’s own, since one could not know what would count as correct or
incorrect instantiations of them. They can only be learnt from those already adept in their
use and aware of these norms (Hamlyn 1978:58-60).

ii) The second problem concerns the other pole, the mature state – Gardner’s ‘end-state’.
We understand this notion well enough in physical contexts like fully-grown hollyhocks
or human bodies. A fully-grown human body is one which can grow no further: it has
reached the limits of its development. The same is true of delphiniums and oak trees.
Like these, the human body can certainly go on changing,but the changes are to do with

7
the maintenance and deterioration of the system, not with its further growth. If we apply
these ideas to the mind, do we want to say that all human beings have mental ceilings -
e.g. in each of Gardner's intelligences - beyond which they cannot progress? This goes
against the grain for many of us. We like to think of our intellectual life as expandable
and deepenable, in principle, in all sorts of directions. True, psychologists like Cyril Burt
have built the notion of mental ceilings into their notion of intelligence, but their views
have been rightly criticized. The claim that we all have individually differing intellectual
limits is both unverifiable and unfalsifiable. As such, it is a metaphysical claim, on a par
with the claims that all historical events are predetermined, or that a personal deity exists.
It is not a scientific claim at all. (White 1998a:29-32)

One answer to this might be that the development of intelligence is unlike physical
development in that here there are no ceilings, simply the potential for endless growth.
Grounds would have to be provided for this claim - which is tantamount to saying that
mental development fails to manifest a feature found in biological development. But if
we leave this on one side, the claim still includes the idea of growth towards states of
relative maturity, even if ceilings are not to be found. It is not clear whether Gardner
would embrace this claim. On the one hand he writes of 'end-state' performances
(1983:64), which suggests finality; on the other, he describes the process of development
as leading to 'exceedingly high levels of competence', which does not.

Whichever view he takes, he still has to say what counts as maturity in the case of the
intelligences. With the oak tree and the human body, we know through the use of our
senses when maturity has occurred: over time we can see that a person is fully grown,
physically speaking, or that an oak tree has reached its full dimensions. What equivalent
is there in the mental realm? How do we know either that people have reached their
mental ceiling or, on the ceiling-less view, that they are more mentally mature than they
were?

We do not just use our senses. We cannot see a person's intellectual maturity as we can
see that he or she is physically fully grown. So how do we tell?

In ordinary life we make all sorts of judgments about people's intellectual or moral
maturity. These judgments tend to be controversial. Some people would understand
intellectual maturity in quiz show terms, as being able to marshal and remember heaps
of facts; others would emphasize depth of understanding; yet others a synoptic grasp of
connexions between different fields; and so on. In moral understanding, what counts as
maturity is similarly contentious. Candidates here might be: having seen the need to
conform to certain absolute rules; sensitivity to others' needs; a philosophical
understanding of what morality is all about; an awareness of the great plurality of moral

8
values and the need to strike sensitive balances among them.

Judgments of mental or moral maturity lack the consensus found in judgments about fully
grown pine trees or badgers. This is because we are in the realm of value judgments
rather than of observable facts: different people apply their own intellectual or moral
values.

Gardner's examples of high levels of development in the intelligences seem to reflect his
own value judgments about what kinds of qualities are important. He starts from
problem-solving skills which are useful and important within a cultural context. He has in
mind the achievements of outstanding poets, composers, religious leaders, politicians,
scientists, novelists and so on. Would everyone agree with him, for instance, that the poet
is the best example of a person whose sensitivity to the meanings and other features of
words has reached a high degree of perfection? Why not the philosopher or propagandist?

In his introduction to the second (1993) edition of Frames of Mind, Gardner backs off
from using only ethically acceptable persons to illustrate the higher realms of the
intelligences. He writes

intelligences by themselves are neither pro-social nor antisocial.

Goethe used his linguistic intelligence for positive ends, Goebbels his
for destructive ones; Stalin and Gandhi both understood other
individuals, but put their interpersonal intelligences to diverse uses.
(1993::xxvi»)

This may cast doubt on whether ‘end-states’ are always achievements highly valued
within a culture. But the more central point is that in the earlier 1983 version of MI
theory (repeated in the 1993 main text) the 'end states' are identified not by observation of
what happens in nature, as with plants or bodies, but by what is held - by Gardner - to be
socially valuable. It is his value judgments, not his empirical discoveries as a scientist,
that are his starting point.

I have tried to show that whether we look towards the beginning or towards the end of the
development process, towards the seed or towards the full flowering, we find apparently
insuperable problems in identifying mental counterparts to physical growth. Since
developmentalist assumptions are central to Gardner's MI theory, the latter is seriously
undermined.

‘susceptibility to encoding in a symbol system’

9
Gardner writes:

following my mentor Nelson Goodman and other authorities, I conceive of a

symbol as any entity (material or abstract) that can denote or refer to any other
entity. On this definition, words, pictures, diagrams, numbers, and a host of other
entities are readily considered symbols (1983:301).

It is important to see how wide the range of Gardner’s symbols is. They include not only
obvious ones like words and mathematical symbols, but also paintings, symphonies,
plays, dances and poems. It is because works of art are symbols in his view that he can
connect many of his intelligences with their own kind of symbolic entities. For instance,
it is not only words which are the symbols associated with linguistic intelligence: this
also contains such symbols as poems.

Gardner states that ‘it may be possible for an intelligence to proceed without its own
symbol system’ (1983:66). He may be thinking here of the personal intelligences (see
also 242). But these, he argues, are dependent on ‘a symbolic code supplied by the
culture’ (ibid), presumably linguistic symbols. Or he may be thinking of bodily-
kinaesthetic intelligence. Although, if works of art can be symbols and mime and dance
are symbols associated with this intelligence, what is one to say about the achievements
of the swimmer, the boxer or the athlete? Are these symbolic insofar as they incorporate
aesthetic elements? Or are they not symbolic entities at all? Gardner does not tell us.

A controversial feature of Gardner’s position is his extension of the notion of a ‘symbol’

to works of art. He writes

In addition to denoting or representing, symbols convey meanings in another

equally important but less often appreciated way. A symbol can convey some
mood, feeling or tone… Thus a painting, whether abstract or representational, can
convey moods of sadness, triumph, anger, or “blueness” (even if the painting itself
is red!). By including this important expressive function within the armament of a
symbol, we are able to talk about the full range of artistic symbols, from
symphonies to square dances, from sculpture to squiggles, all of which have
potential for expressing such connotative meanings. (1983: 301)

It looks as if Gardner is saying that not all symbols are like words. A word is a symbol in
that it stands for something outside itself. The English word ‘cat’ and the French word
‘chat’ refer to the same kind of thing in the world, namely cats. But some symbols -

10
some works of art - do not denote or represent: they convey or express feelings.

The difficulty with this is that while works of art can be expressive of emotion, it is hard
to see why they should be called ‘symbols’ for that reason. What are they symbolizing?
One can understand the notion readily enough when talking about words, flags or
communion wine. In each of these cases one can draw a distinction between the symbol
and what it is a symbol of: cats, America, the blood of Christ. If a song is a symbol in the
same way, what is the thing symbolized which lies outside it? Suppose it expresses
sadness. Is sadness what it is symbolizing?

Gardner does not say so. It would be hard, in any case, to argue that it is symbolizing this
– or, indeed, that works of art symbolize anything – even though some of them, portraits,
for instance, represent something outside themselves. We can understand symbolizing
well enough by reference to words and flags. Here a symbol not only picks out something
else. It is also replaceable by equivalent symbols, those with the same function. Words
are translatable into other languages; a flag replaceable by a statue. But a Rembrandt
portrait or Mozart piano concerto cannot be replaced by symbols with an equivalent
function: works of art are unique. If they are said to be symbols, we cannot understand
what this means by reference to what we know of symbols in uncomplicated cases. The
use of the term in Gardner is obscure. If in an artistic context ‘symbolising’ means no
more than ‘expressing feeling’, the term is redundant. In addition, ‘symbol’ now comes to
have a different meaning in the arts from what it has in language and in mathematical
thinking. On the other hand, if Gardner’s ‘symbolizing’ is not equivalent to ‘expressing
feeling’, it is difficult to see what sense can be attached to it. We do not know what a
work is supposed to symbolize or in what way its being a symbol is like or unlike
symbolism in the unproblematic sense. The term – and the theory of MI that depends on
it – become radically obscure.

Without going through all the other criteria, a word about two of them.

‘ the potential isolation of the area by brain damage’

The criteria to do with development and with symbol are central items on Gardner’s list.
This can be seen if one tries to imagine their absence. I shall come back to the centrality
of the symbol criterion later. Meanwhile let us imagine the exclusion of the development
criterion. Suppose we take what appears to be the weightiest of the other criteria: ‘the
potential isolation of the area by brain damage’. And let us take it that there are localized
areas of function within the brain. If one part of the brain is damaged, one’s sight is
impaired, if another, one’s ability to move one’s left hand, or feel pain, or talk, or
understand speech. What this shows is that certain physiological necessary conditions of

11
exercizing these capacities are absent. It does not help to indicate the existence of
separate 'intelligences'. It is well known that language ability is impaired through brain
injury to parts of the left hemisphere of the cerebral cortex. But the injury could in
principle impair wired-in abilities implicated not only in language use but in all sorts of
other things as well; and there does indeed seem to be empirical evidence that this is the
case (Richardson 1999: 85-8). The capacities in question are not those of a language
module but of ‘much more general and lower-grade functions’ (87).

Given his developmentalism, one can understand why Gardner should look to brain
localization in order to identify intelligences, for he has to provide an account of the
'seed' which is to unfold into its mature form, and this seed has to be part of our original,
biologically given, constitution. But the kinds of function picked out by brain localization
research do not have the power, as far as I can see, to grow into more developed forms. I
am indeed born with the power of vision or the power to move my thumbs, but although
various forms of socialization are built on these abilities, the latter do not themselves
unfold into maturer versions of themselves.

‘the existence, in an area, of idiots savants, prodigies and other exceptional

individuals’

Gardner invokes the existence of idiots savants to support his theory, but what I know of
them does not lead me to think of them as intelligent. Well known recent examples
include an 11-year-old London boy who can draw complicated buildings perfectly having
just seen them; a 23-year-old man who can play piano pieces perfectly having heard them
only three times; and a young man who can tell you the day of the week of any date
presented to him. All these cases are of subnormal mental ability. What they all have in
common is a mechanical facility, one which lacks the flexibility of adapting means to
ends found in intelligent behaviour.

Prodigies only support Gardner’s case if there is good evidence that their talents are
innate. But the evidence there is seems to point to acquired abilities (Howe 1997: 131-2)

Conclusion

It would be natural to think that the ‘criteria’ against which one measures candidate
intelligences that have survived the first cut are all straightforwardly applicable – in the
sense that it is an empirical task (although perhaps a complicated one) to look at the
relevant facts and come to a judgment. But this is not so. The criteria to do with

12
development and with symbols presuppose the truth of theories – one in psychology, the
other in aesthetics – which, once subjected to philosophical critique, turn out to be
untenable. And this undermines the viability of MI theory as a whole.

How are the criteria to be applied?

How does one use the criteria to pick out intelligences? If they are all necessary
conditions, each has to be met before we can say that an intelligence exists. Although
some of them seem to be necessary - to judge by remarks like 'an intelligence must also
be susceptible to encoding in a symbol system' (Gardner, 1990:933), in his original work
Gardner makes it clear that not all have to be satisfied (Gardner, 1983:62). In places, the
demand is more stringent. In his 1990 discussion of how he came to pick out his
intelligences, he writes that 'only those candidate intelligences that satisfied all or a
majority of the criteria were selected as bona fide intelligences' (Gardner, 1990:932). If
this is to be taken literally, then if five or more of the eight criteria listed are met, a
candidate automatically passes the test. But Frames of Mind states that there is no
'algorithm for the selection of an intelligence, such that any trained researcher could
determine whether a candidate intelligence met the appropriate criteria’ (p.63). Rather,
Gardner goes on:

At present, however, it must be admitted that the selection (or rejection) of a

candidate intelligence is reminiscent more of an artistic judgment than of a
scientific assessment, (p.63)

The identification of intelligences appears, then to be a subjective matter, depending on

the particular weightings that Gardner gives to different criteria in different cases.

It is worth dwelling on this point. Gardner sees it as a special virtue of his theory, which
differentiates it from rival accounts, that it is scientifically based. He writes

There have, of course, been many efforts to nominate and detail essential
intelligences, ranging from the medieval trivium and quadrivium to the
psychologist Larry Gross’s list of five modes of communication (lexical, social-
gestural, iconic, logico-mathematical, and musical), the philosopher Paul Hirst’s
list of seven forms of knowledge (mathematics, physical sciences, interpersonal
understanding, religion, literature and the fine arts, morals, and philosophy). On
an a priori basis, there is nothing wrong with these classifications; and, indeed,
they may prove critical for certain purposes. The very difficulty with these lists,

13
 however, is that they are a priori – an effort by a reflective individual (or a
culture) to devise meaningful distinctions among types of knowledge. What I am
calling for are sets of intelligences which meet certain biological and
psychological specifications. In the end, the search for an empirically grounded
set of faculties may fail; and then we may have to rely once more on a priori
schemes, such as Hirst’s. But the effort should be made to find a firmer foundation
for our favourite faculties. (1983: 61-2).

In saying that selecting intelligences is more like making an artistic judgment than a
scientific assessment, Gardner thus seems to be contradicting himself. The non-empirical
nature of his theory has also been shown above. We have seen how the ‘first cut’
selection of the intelligences is not based on empirical investigation of what different
societies have held to be valuable; and that the ‘criteria’ depend on theories in psychology
and aesthetics which themselves are not empirically founded.

Gardner has replied to the charge I made originally in White 1998b, based on the same
quotation about artistic judgment, that his choice of intelligences is subjective. He wrote

White correctly notes that my original list depended on the judgment of a single
analyst, who made his data available to others. However, White is naïve if he
believes that science begins in any other way. (Gardner 1998)

What he may have in mind is the Popperian point that science begins with conjectures.
But not all conjectures eventuate in science. Some may be empirically untestable – the
conjecture, for instance, that every event is predetermined. A fundamental question about
MI theory is whether it is empirically testable. Because it is not clear when a candidate
intelligence passes or does not pass the ‘criteria’ test, it is uncertain under what
conditions it might be refuted. Empirically refuted, that is. If I am right in what I said
about development and symbols, it may be that MI theory is refutable on philosophical
grounds. But Gardner needs not a priori, but empirical, refutability.

Why these criteria?

A further - and surely fundamental - question is: how does Gardner justify using the
particular criteria he lists to pick out intelligences?

I have not been able to find any answer in his writings. Whenever he introduces the
criteria, they are each spelt out in some detail, but there is no account of why these ones

14
have been employed and not others. In a properly scientific theory one would expect a
rationale for this.

MI theory in biographical perspective

That is not to say that no explanation can be given. I believe there is a reason. But it is not
a reason in the sense Gardner needs. What one can do is tell a plausible historical story
about how he came to put weight on the criteria. This would be an explanation of how
they have come about, but not a justification of them. And it is a justification that Gardner
needs to make good his claim that his is an objective, scientific theory. In what follows I
shall concentrate on the same two major criteria examined above – to do with
development and with symbol systems. As we shall see, the historical discussion will
throw light not only on the criteria, but more broadly on MI theory as a whole. It will
help to make sense of it. It will also bring us back to the vital importance of Gardner’s
‘first cut’.

As far as I can judge, the historical story runs like this. In the 1960s Gardner began his
career as a developmental psychologist, profoundly influenced by Piaget as well as by
structuralist thinkers in other fields, notably Lévi-Strauss in anthropology.

The structuralists are distinguished first and foremost by their ardent, powerfully
held conviction that there is a structure underlying all human behavior and mental
functioning…(1972:10)

The young Gardner was also ‘a serious pianist and enthusiastically involved with other
arts as well’ (2003:1).

No surprise, then, that when I first began to study developmental psychology, I

was soon struck by certain limitations in the field. The child was seen by nearly all
researchers as an exclusively rational creature, a problem-solver – in fact, a
scientist in knickers….While a first-year graduate student, I elected to direct my
own research toward a developmental psychology of the arts. (1982: xii)

He goes on

The structuralist approach to the mind has limitations.…Though creative thought

has not escaped their attention, each of the major cognitive structuralists views

15
 the options of human thought as in some way preordained, limited in advance.
This makes their work especially problematic for a study of the mind where the
major focus falls on innovation and creation, as in the fashioning of original
works of art.

To my mind the limitation implicit in the standard structuralist stance can be

circumscribed by a recognition of one special feature of human thought – its
ability to create and sponsor commerce through the use of various kinds of
symbol systems. These symbol systems – these codes of meaning – are the
vehicles through which thought takes place: by their very nature they are
creative, open systems. Through the use of symbols the human mind, operating
according to structuralist principles, can revise, transform and re-create wholly
fresh products, systems, and even worlds of meaning. (op.cit.:4-5)

These various quotations indicate the centrality of theories of development and of symbol
systems in Gardner’s pre-1983 thinking. Much of his published work in this period was
about the application of developmental psychology to the arts. He describes his 1973
book The Arts and Human Development as

fleshing out the picture of development proposed by Piaget, extending it to the

vast majority of people who are not scientists … but who nonetheless participate
in a significant way in complex intellectual activities. (1973:vii)

Key to this extension of Piagetian ideas to the arts is the notion of a ‘symbol’. Here
Gardner was influenced by the aesthetician Nelson Goodman, co-founder with himself of
Project Zero at Harvard Graduate School of Education in 1967. Goodman, in the tradition
of Ernst Cassirer and Suzanne Langer, saw works of art as a whole as symbols and also
as containing symbols within themselves. Different arts, he held, have their own symbol
schemes – hence the title of his Languages of Art (Goodman 1968). Sometimes artistic
symbols have a denotating function, as words do, in that they stand for something outside
themselves. In fact words constitute some works of art; other denotating symbols are
portraits and drawings from nature. Other artistic symbols ‘exemplify’ rather than denote.
Just as a sample of household paint exemplifies the properties of the paint itself, so a
piece of music, in its sad qualities, exemplifies sadness generally – not literally but
metaphorically. Understanding a work of art is not a matter of appreciation or aesthetic
experience, but of interpreting correctly what and how it symbolizes. The arts, for
Goodman, are forms of knowledge.

Gardner’s early intellectual biography throws light on his Frames of Mind, especially the
first five intelligences: linguistic, musical, logico-mathematical, spatial and bodily-

16
kinaesthetic. Of these, logico-mathematical intelligence is related particularly to
mathematics and science and its treatment follows Piaget’s scheme quite closely. The
other four intelligences reflect Gardner’s work in extending Piagetian developmentalism
into the arts: poetry is prominent in the chapter on linguistic intelligence, music in that
the musical chapter, the visual arts in the spatial, mime and dance in the bodily-
kinaesthetic.

Piaget’s and Goodman’s theories are examples of developmentalism and the symbol
theory of art respectively, both of which were criticized above. In addition, there are
further conceptual problems damaging to these two theories taken specifically. These
have been explored by David Hamlyn in the case of Piaget (Hamlyn 1967, 1978) and
Roger Scruton in the case of Goodman (Scruton 1974).

Until 1979 Gardner’s work extended Piagetian thinking into the arts. By 1983 it
broadened again, now into an overall theory of human intelligences.

I shall explore why in the next section. In this present section I have argued that if we ask
why he used the criteria he did, we look in vain for a justification from Gardner himself,
but find a likely explanation in the historical progress of his ideas. I have explored this
for the criteria to do with symbols and development.

The van Leer Project

A crucial turning point for Gardner came in 1979, when he moved from his long-standing
project on the development of artistic competences to an all–embracing theory of
intellectual development. Why?

The answer, as he indicates in Gardner 2003, has to do with the Harvard Project on
Human Potential funded by the Bernard van Leer Foundation in 1979. The Foundation

asked the Harvard Graduate School of Education to assess the state of scientific
knowledge concerning human potential and its realization and to summarize the
findings in a form that would assist educational policy and practice throughout the
world. (Gardner 1983: x)

Gardner’s task in the interdisciplinary team was to look at psychological, as distinct from
philosophical and anthropological, considerations. Frames of Mind was the first

17
publication from the team. Unlike the second volume, Of Human Potential, Israel
Scheffler’s (1985) philosophical investigation of this notion, Gardner’s book did not look
globally at the topic, since it focused on ‘human intellectual potential’ (1983:x) (my
italics).

As suggested above, the first five intelligences in this book drew on Gardner’s pre-1979
work in the Piagetian and Goodmanian traditions and areas of interest. The other two
were the personal intelligences. It is understandable that Gardner should wish to include
other areas of interest than mathematics, science and the arts. The van Leer remit wanted
something more comprehensive. Strictly, a general study of human potential – insofar as
one can attach a defensible sense to the notion (White 1986) – is unrestricted by subject-
matter. Human beings can possess capabilities of innumerable sorts, not only intellectual
ones connected with the pursuit of knowledge, but also practical ones to do with
threading needles, cutting one’s toenails, running merchant banks. Gardner did not take
on this whole gamut. He restricted himself to intellectual areas, in the sense of those
concerned with the pursuit of knowledge. (I bypass his inclusion of artisans and others in
his discussion of bodily-kinaesthetic intelligence – see above). Having agreed, as his
contribution to the van Leer project, to look at human intellectual potential, he had to
determine what further intellectual areas should be considered beyond the sciences,
mathematics and the arts. The result was the two personal intelligences – those to do with
understanding oneself and understanding other people.

In answer to his own question ‘why have I incorporated personal intelligences in my

survey?’, Gardner says

Chiefly because I feel that these forms of knowledge are of tremendous importance
in many, if not all, societies in the world – forms that have, however, tended to be
ignored or minimized by nearly all students of cognition. (1983:241)

This is revealing. It shows him making another ‘first cut’. The personal intelligences pass
this test because of their huge social importance. Gardner then sees how far they pass the
second, ‘criteria’, test. In one way, the first test is sufficient. If you are reviewing the full
range of forms of understanding, and so far your work has included only scientific,
mathematical and artistic understanding, it will not take long to light on self-
understanding and understanding of others as further significant areas, given, as Gardner
says, their importance in human life.

He may reply, perhaps, that his task was not to identify forms of knowledge in an a priori
way – after all, Paul Hirst and others had done that. He needed empirical evidence for
them. This meant seeing how well the personal intelligences meet the second-cut criteria.

18
Contra Gardner, they do not meet them very well. He devotes a section to ‘the
development of the personal intelligences’ from infancy through to maturity (1983: 243-
253). The evidence he produces is of changes in understanding - which becomes more
sophisticated and discriminating in various ways. It is not evidence of change which is
also an unfolding. As Gardner for the most part treats ‘the development of personal
knowledge as a relatively natural process’ (1983:253), he radically underplays the role of
young children’s mentors, especially their parents, in inducting them into the language,
experiences, ethical involvement and folk-psychological insight necessary for this kind of
understanding. Gardner’s preconception that development must occur appears to blind
him to what most people would see as an obvious explanation.

The symbol test does no better. Gardner admits that ‘one does not ordinarily think of
forms of personal knowledge as being encoded in public symbol systems’, and falls back
on the true, but lame, thought that these forms of knowledge depend on ‘a symbolic code
supplied by the culture’ (1983:242), ie presumably everyday words in the language. Since
a person could hardly be said to possess any intellectual competence unless he or she
were a language-user, Gardner’s point cuts little ice.

In addition, what he says about brain lesion evidence points to changes in mood and
emotion, eg depression, following injury to the frontal lobes, but not, as far as I can see,
to changes in self- or other understanding. He himself says that positive evidence is
sparse with regard to the criteria to do with evolutionary evidence, exceptional
individuals, experimental psychology and psychological testing (1983:242). Despite all
this discouraging news, the two intelligences get their diplomas.

This prompts the thought: once they were past the first cut, were they not going to get
them anyway? Gardner has always emphasized that little turned for him during the van
Leer Project on calling his multiple competences ‘intelligences’.

I don’t remember when it happened but at a certain moment, I decided to call

these faculties “multiple intelligences” rather than abilities or gifts. This
seemingly minor lexical substitution proved very important; I am quite confident
that if I had written a book called “Seven Talents” it would not have received the
attention that Frames of Mind received. (2003:4)

Another synonym in play at that time was ‘forms of knowledge’.

..nothing much turns on the particular use of this term [‘intelligences’], and I
would be satisfied to substitute such phrases as “intellectual competences,”

19
 “thought processes,” cognitive capacities,” “cognitive skills,” “forms of
knowledge,” or any other cognate mentalistic terminology. (1983:284)

We saw above that Gardner’s reason for including the personal intelligences was that

I feel that these forms of knowledge are of tremendous importance in many, if not
all, societies in the world… (1983:241)

We also saw above that he included Paul Hirst’s list of seven forms of knowledge among
the ‘many efforts to nominate and detail essential intelligences’ and that if his own
empirical approach to demarcating intelligences should fail, ‘then we may have to rely
once more on a priori schemes, such as Hirst’s’ (1983:61-2).

Once Gardner saw that the personal competences were tremendously important forms of
knowledge (ie intelligences), it is hard to see how they could be excluded from his final
list. On this line of thinking, the ‘first cut’ looks to be the crucial one.

A further question arises here about how Gardner conceived his project in this 1979-83
period. If his intelligences are in the same ball park as Hirst’s forms of knowledge – and
indeed as ‘the medieval trivium and quadrivium’ (ibid.), can they still be equated with
abilities or talents? From the former point of view, they come out as ways of categorizing
the realm of intellectual phenomena; from the latter, as ways of categorizing individuals’
intellectual competences. The first classification is extra-individual: it is of
epistemological phenomena in the social world. The second is intra-individual - of
attributes of persons.

For Gardner at this time the two ways of classifying were linked. He saw his theory as
bridging the – bio-psychological - world of individual nervous systems and the –
epistemological or anthropological - world of social forms. Symbols have a central role
in this.

..there is no ready way to build a bridge directly between these two bodies of
information [biology and anthropology], their vocabularies, their frames of
reference are too disparate….The domain of symbols, as it has been constituted
by scholars, is ideally suited to help span the gap between the aforementioned
entities. (1983:300)

Outside Gardner’s theory, the two classifications can be kept apart. Paul Hirst, for
instance, saw himself as doing epistemology, not psychology. His theory is about how
knowledge is to be logically carved up, not about the kinds of intellectual abilities

20
belonging to individuals.

For Gardner, the two spheres are inseparably connected. This is implicit in his
developmentalism and his symbol theory: abilities unfold from seeds within the nervous
system towards mature end-states found in different intellectual activities; and it is
through the acquisition of symbols that these end-states are those of the highest flights of
creative activity. Because of this inseparable connexion, studying one pole of the process
throws light on steps leading to the other. The bio-psychological study of individuals is a
key to the social/epistemological world of the disciplines; and vice-versa.

Conclusion

It has become clear that the requirements of the van Leer project allowed Gardner to
expand from the limited theory of artistic development on which he had previously
concentrated to a fuller account of the development of human intellectual competences as
a whole. In doing so, he was able to retain the master-ambition which had motivated his
work from his earliest days as a structuralist, bringing Piagetian insights into harmony
with those of Lévi-Strauss: the desire to link biology and anthropology, to show that they
are part of the same system.

MI Since 1983

Since 1983 there have been several modifications of MI theory.

[a] The original seven intelligences have now been extended to include ‘the naturalist
intelligence’ and – possibly – ‘existential intelligence’ (1999a: ch4). Naturalist
intelligence is picked out by reference to a valued social role found across many cultures:
people expert in recognizing and classifying the varieties of plants and animals in their
environment. This is the ‘first cut’. Naturalist intelligence is then shown to satisfy all or
most of the ‘criteria’.

There are by now familiar points to be made about how well the criteria are met, not least
those to do with development and with symbols. But a prior question is why the new
naturalist intelligence came to be proposed in the first place. Gardner tells us that ‘those
valued human cognitions that I previously had to ignore or smuggle in under spatial or
logico-mathematical intelligence deserve to be gathered under a single, recognized rubric’
(1999a:52). This seems to imply that, having reviewed the full gamut of intellectual

21
activities, he realized that the taxonomic aspects of biology had been given short shrift in
his original scheme.

This thought is reinforced by what he says in the same chapter on possible forms of
spiritual intelligence and of existential intelligence – to do with ‘big questions’ about
one’s place in the cosmos, the significance of life and death, the experience of personal
love and of artistic experience - as the strongest candidate among these (pp53-65).
Religious and philosophical thinking are also parts of the intellectual world; and these,
too, were ill-represented in the 1983 scheme.

All this lends strength to the suggestion that what powers MI theory is the attempt to
identify all major divisions of the intellectual life (taking the arts as always to be forms of
knowledge). Crucial to the theory, as we have seen, is the ‘first cut’.

[b] Since 1983, too, Gardner has become bolder about the significance of MI theory.
What began as a response to an external funder’s request – the extension of a pre-existing
interest in development in the arts as well as in Piagetian areas into a more global survey
of ‘human potential’ – has generated

a new definition of human nature, cognitively speaking. Whereas Socrates saw

man as the rational animal and Freud stressed the irrationality of human beings, I
(with due tentativeness) have described human beings as those organisms who
possess a basic set of seven, eight, or a dozen intelligences (1999a:44).

This, it seems to me, is to write a prescription about desirable intellectual attainments into
a description of human nature as a whole. In any case, why make intellectual activity the
defining characteristic of our human nature? Human beings are into all sorts of other
things than the pursuit of knowledge and the arts. So why highlight these?

[c] A third – related - departure since 1983 has been Gardner’s distinction between
‘intelligence’ and ‘domain’ (1999a:82). The former is ‘a biopsychological potential that is
ours by virtue of our species membership’. The latter is a ‘socially constructed human
endeavor’, for example ‘physics, cooking, chess, constitutional law, and rap music’. It is
‘characterized by a specific symbol system’. Gardner says he could have made this
distinction more carefully in 1983. Readers would then have seen more clearly that
several intelligences can be applied in the same domain, and the same intelligence in
many domains.

This move detaches from each other the two dimensions – the biological and the social –
which Gardner tried to hold together through his career. It makes MI theory

22
unintelligible. For it has always been part of the concept of an intelligence that it is an
ability that develops from a physiological origin towards an end-state belonging to a
valued social activity. Poetry, music, the visual arts, dance, mathematics, logic, sport –
the loci of the 1983 end-states – are all social constructions. Similarly, the idea of an
intelligence was originally founded partly on the thought that symbols are bridges
between the biological and the social. The 1999 version separates the previously
inseparable and puts symbols and end-states firmly on the side of the social – as attached
to domains rather than intelligences. At the same time, the ‘criteria’, which remain
unchanged from 1983, include reference to both symbols and end-states among the
distinguishing features of intelligences. This is why the 1999 version of MI theory is
unintelligible.

MI and education

Until the van Leer project Gardner was a psychologist, not an educationalist. But he had
to adhere to the van Leer request that the Harvard team summarize its findings about
human potential ‘in a form that would assist educational policy and practice throughout
the world’. (Gardner 1983: x). This was because the van Leer Foundation was an
international non-profit making institution dedicated to helping disadvantaged children to
realize their potential (Scheffler 1985:ix). In Frames of Mind Gardner

touched on some educational implications of the theory in the concluding

chapters. This decision turned out to be another crucial point because it was
educators, rather than psychologists, who found the theory of most interest.
(2003:4)

Since 1983 MI theory has had a huge influence on educational reform, especially school
improvement, across the world. It has affected its views about pupils and their aptitudes,
methods of learning and teaching, and curriculum content. If the argument of this essay is
correct, all this has been built on flaky theory.

[a] Gardner holds that while nearly all children possess all the intelligences to some
degree, some of them have particular aptitudes in one or more of them. ‘My own belief is
that one could assess an individual’s intellectual potentials quite early in life, perhaps
even in infancy’ (1983:385)

23
It is not surprising that ideas like these have – not intentionally - encouraged educational
policies and practices to do with selection, specialisation, individualisation of learning,
and assessment. But if the intelligences are not part of human nature but wobbly
constructions on the part of their author, educators should treat them with caution. There
may or may not be good grounds for personalised learning and other policies, but if they
exist they must come from elsewhere. That teachers often need to vary the way they teach
according to what best motivates particular pupils has been part of pedagogy for
centuries; there is no good reason for confining this notion within the ‘intelligences’
framework.

[b] There is abundant evidence that MI theory has been influential in reducing the low
self-esteem of pupils who see themselves as stupid or thick, where this kind of judgment
derives from conventional ideas of general intelligence based on IQ. The idea that
intelligence is not necessarily tied to prowess in logical, mathematical and linguistic tasks
but can be displayed across a variety of fields is true – as our opening quotation from
Gilbert Ryle illustrates. But the idea is not necessarily tied to MI theory (White 1998a:3-
4).

[c] One reason why MI theory has been so influential may be its basis in supposedly
discrete forms of intellectual activity – in Gardner’s broad use of the term to embrace not
only disciplines based on the pursuit of truth like biology and mathematics, but also the
arts and athletics. With some exceptions, the areas it covers are close to those in a
traditional so-called ‘liberal education’ based mainly on initiation into all the main areas
of knowledge, to be pursued largely for their own sake. The addition of naturalist
intelligence and (possibly) existential intelligence have made the fit even closer, seeing
the affinities of these areas with biology and with work of a philosophical/religious sort.

On the whole, Gardner has refrained from deriving curricular consequences from MI
theory. His writings on what the content of education should be show that the type of
schooling he favors is in the ‘liberal education’ tradition.

Education in our time should provide the basis for enhanced understanding of our
several worlds – the physical world, the biological world, the world of human
beings, the world of human artifacts, and the world of the self. (1999b:158)

He also thinks this understanding should be largely for intrinsic ends.’I favor …the
pursuit of knowledge for its own sake over the obeisance to utility’ (1999b:39). This
locates him firmly within the ‘liberal education’ camp, along with - in Britain - (the
early) Paul Hirst, Richard Peters, Roger Scruton and others.

24
It is not surprising that Gardner’s curricular ideas dovetail with his ideas of the
intelligences, even if this was not his original intention. For the ‘liberal education’
tradition and MI theory share the same starting point. They both assume the importance
in human life of intellectual activities pursued largely for their own sake.

It is not surprising that educators reacting against recent utilitarian tendencies in

schooling and looking for a more humane alternative have been attracted by MI theory,
given its closeness to traditional ‘liberal education’. But the latter idea is not necessarily
tied to MI. Hirst, Peters and others have argued for it on quite other grounds (Hirst 1974,
Peters 1966: ch5).

What is more, ‘liberal education’ – in this sense of intellectual learning for its own sake -
itself needs justification. The reasons I have seen in favour of this do not hold water
(White 1982:ch2). There is a danger that in basing children’s schooling on it we are
imposing a life ideal on them – to do with the pursuit of truth and beauty for their own
sake – which we, as intellectually inclined people, may find personally appealing, but
which, after all, is only one of many possible life ideals. There is a good case for a
broader view of educational aims that, while celebrating intrinsic intellectual aims, also
embraces other perspectives on a worthwhile human life, thus leaving young people less
confined in deciding the paths they wish to take.

Note

I am most grateful to the following UK teachers for information about how MI theory has
been applied in their schools: Margaret Grant, Deputy Headteacher of Broughton Hall
School, Liverpool; and James McAleese, of Richard Hale School, Hertford.

Bibliography

Gardner, H. (1972) The Quest for Mind London: Coventure

---------- (1973) The Arts and Human Development New York: John
Wiley

---------- (1982) Art, Mind and Brain New York: Basic Books

---------- (1983) Frames of Mind: the theory of multiple intelligences

25
 London: Heinemann

---------- (1990) ‘The theory of multiple intelligences’ in Entwistle, N.

(ed) Handbook of Educational Ideas and Practices
London: Routledge

---------- (1993) Frames of Mind: the theory of multiple intelligences

(Second edition) London: Heinemann

---------- (1998) ‘An intelligent way to progress’ The Independent

19.3.98

---------- (1999a) Intelligence Reframed: multiple intelligences for the

21st century New York: Basic Books

---------- (1999b) The Disciplined Mind New York: Simon and Shuster

---------- (2003) ‘Multiple Intelligences after Twenty Years’ Paper

presented at the American Educational Research Association,
Chicago, Illinois, April 21 2003
http://www.pz.harvard.edu/PIs/HG_MI_after_20_years.pdf.

Goodman, N. (1968) Languages of Art Indianapolis: Hackett

Hamlyn, D. (1967) ‘Logical and psychological aspects of learning’ in Peters,

R.S. (ed) The Concept of Education London: Routledge
and Kegan Paul

---------- (1978) Experience and the Growth of Understanding London:

Routledge and Kegan Paul

Hirst, P.H. (1974) ‘Liberal Education and the Nature of Knowledge’ in

Knowledge and the C urriculum London: Routledge and
Kegan Paul

Howe, M.J.A. (1997) The IQ in Question London: Sage

Peters, R.S. (1966) Ethics and Education London: Allen and Unwin

Richardson, K. (1999) The Making of Intelligence London: Weidenfeld and Nicolson

26
Ryle, G. (1949) The Concept of Mind London: Hutchinson

Scheffler, I. (1985) Of Human Potential London: Routledge and Kegan Paul

Scruton, R. (1974) Art and Imagination London: Methuen

White, J. (1982) The Aims of Education Restated London: Routledge and

Kegan Paul

--------- (1986) ‘On reconstructing the notion of human potential’ Journal

of Philosophy of Education Vol 20, No 1

--------- (1998a) Do Howard Gardner’s multiple intelligences add up?

London: Institute of Education University of London

--------- (1998b) ‘Intelligence guru on a sticky wicket’ The Independent

19.2.98

27